- Email: [email protected]
Decreased health care use among patients with silent myocardial ischemia
Journal of Psychosomatic Research 48 (2000) 479–484
Decreased health care use among patients with silent myocardial ischemia: support for a generalized rather than cardiac-specific silence Mark Lumleya,*, Laurenn Rowlanda, Tracey Torosianb, Adam Banka, Mark Kettererb, Sol Pickardc a Department of Psychology, Wayne State University, 71 West Warren Avenue, Detroit, MI 48202, USA Department of Psychiatry, Henry Ford Health System, 2799 West Grand Boulevard, Detroit, MI 48202, USA c Division of Cardiovascular Medicine, Henry Ford Health System, 2799 West Grand Boulevard, Detroit, MI 48202, USA Received 30 September 1998; accepted 1 December 1999 b
Abstract Objective: The absence of angina among patients with silent myocardial ischemia (SMI) may be a cardiac phenomenon or may reflect a generalized lack of bodily awareness and symptom reporting. We tested the hypothesis that the silence is generalized, and, therefore, that patients with SMI would make fewer health care visits for noncardiac/chest-pain problems than patients with symptomatic ischemia. Methods: We counted all out-patient visits to our medical system for the prior 18 months for 95 patients who demonstrated ischemia during treadmill exercise testing and subsequent nuclear scanning: 62 of the patients had SMI during exercise, and 33 of
the patients had symptomatic ischemia. Results: Patients with SMI made were significantly less likely to have sought emergency care or primary care and had significantly fewer primary care visits than patients with symptomatic ischemia. Group differences remained after controlling for demographics and health status variables. The two groups did not differ on utilization of specialty care. Conclusion: The reduced use of emergency and primary care among patients with SMI suggests that they have a generalized rather than cardiac-specific reduction in somatic awareness and/or symptom reporting. 2000 Elsevier Science Inc. All rights reserved.
Keywords: Silent myocardial ischemia; Asymptomatic; Health care utilization; Chest pain; Somatic awareness
Introduction Coronary artery disease (CAD) results in myocardial ischemia when the oxygen supply is insufficient to meet myocardial demands. Ischemia may be accompanied by angina pectoris or other symptoms of pain and discomfort, but the ischemia is often silent or asymptomatic. A paradigm sometimes employed to study the correlates of silent myocardial ischemia (SMI) is to identify patients who exhibit ischemia during treadmill exercise testing, and then to compare the subsample of patients who report no symptoms during exercise-induced ischemia (i.e., who have SMI) with the subsample who report angina during ischemia (i.e., symptomatic ischemic patients). Studies have identified both biological and psychosocial correlates of SMI. Among the biological correlates, SMI has been linked to a lower heart rate at the onset * Corresponding author. Tel.: 313-577-2838; fax: 313-577-7636. E-mail address: [email protected] (M. Lumley).
of ischemia [1,2] and with smaller areas of ischemic tissue [3–5]. SMI also has been associated with specific regions of ischemic myocardium [6–8], diabetic neuropathy [9,10], the use of antianginal medications [11], and cigarette smoking [12]. Other studies have identified psychological correlates of SMI. Compared to patients with symptomatic ischemia, patients with SMI have been found to report less depression, anxiety, somatic awareness, illness seriousness, and sensitivity to pain and discomfort. They also report more marital satisfaction, type A behavior, and internal locus of control [13–20]. We have found that patients with SMI report a greater tendency to control their anger, and they prefer an externally oriented rather than an introspective style of thinking [21]. It remains unclear, however, whether the silence associated with SMI is limited to cardiac phenomena, or whether it reflects a more generalized somatic unawareness and underreporting of bodily symptoms. In a key review, Barsky and colleagues [22] raised this issue by
0022-3999/00/$ – see front matter 2000 Elsevier Science Inc. All rights reserved. PII: S0022-3999(00)00101-X
480
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
asking, “Is the person or the event silent?” Some of the cardiac correlates of SMI just noted (i.e., size and location of ischemic region, heart rate at ischemia onset, and use of antianginal medications) suggest that the silence is cardiac-specific—a function of cardiac tissue or innervation. In contrast, the findings associated with personality, attitudes, emotional style, and pain reporting suggest that the lack of symptoms is global or generalized—a function of the whole person. The available literature does not clearly answer the question of whether SMI is due to the person or the event. The cardiac correlates of SMI have been inconsistent. Some studies did not find lower heart rates at the onset of silent ischemia [21,23], and other studies found no associations between the size of ischemic myocardium and angina [21,24,25]. The psychological findings also have limitations. They have relied almost exclusively on self-reports via questionnaires, which, although purporting to assess different constructs, probably reflect a much smaller number of personality or emotional dimensions. For example, reports on many questionnaires are strongly influenced by a response tendency termed “negative affectivity,” which refers to a person’s disposition to experience and report negative aspects of oneself, including both emotional and physical symptoms [26,27]. Thus, those who experience SMI may simply be those who deny problems or have low levels of negative affectivity. We believe that a more valid test of the etiology of SMI requires measures that are not self-reported— measures such as behavioral indices. In this study, we examine the behavior of health care utilization. We reasoned that, if SMI is a generalized rather than cardiacspecific phenomenon, then patients with SMI would experience and complain of fewer bodily symptoms. This would result in patients with SMI being less likely to seek medical care for various somatic problems, particularly health problems unrelated to cardiac concerns and chest pain. In contrast, if SMI is cardiac-specific, then patients with SMI would be expected not to differ from symptomatic patients in their use of medical care for noncardiac/non–chest-pain problems. Thus, we compared patients who experienced SMI during treadmill exercise with patients who experienced symptomatic ischemia on the frequency with which they utilized various types of health care, including emergency, primary, specialty, and cardiology types of care. For each health care type (except cardiology), we excluded health care contacts that were related to cardiac problems or angina, thus ruling out the possibility that the decreased health care utilization was due solely to cardiac events and symptoms. Methods Participants Consecutive adults (n ⫽ 317) who had treadmill exercise testing with myocardial perfusion imaging were
asked to participate as they presented for testing; 275 (86.8%) agreed and provided informed consent. Of these, 17 were excluded due to cognitive impairment (n ⫽ 15) or illiteracy (n ⫽ 2). Of the remaining 258 patients, 112 (43.4%) were positive for exercise-induced cardiac ischemia (defined in what follows) and, of these patients, 95 (84.8%) were included because our records indicated that they received other out-patient care within our medical system during the prior 18 months. All of the patients in the final sample had private or public medical insurance, and most (n ⫽ 74, 78%) were insured by a managed care corporation affiliated with our health care system and which typically limited services to our system (thus providing us with patients’ complete utilization records). Most patients were referred for treadmill testing with nuclear scanning because they had experienced symptoms considered to be potentially indicative of coronary artery disease (CAD). Also, most patients, including those without symptoms, had other indications for testing, including elevated cardiac risk factors, abnormal or inconclusive resting electrocardiograms (ECGs), or presurgical clearance. Elsewhere, we reported on other characteristics from this cohort of patients, including: psychological correlates of SMI [21]; unrecognized myocardial infarction [28]; noncardiac chest pain [29]; and alexithymia [30]. Procedures Patients’ cognitive status was tested [31], they were interviewed and completed questionnaires for other studies, and were tested for myocardial ischemia. Before exercise, patients received 8.0 mCi of technetium-99msestamibi (Tc-setamibi), and baseline resting tomographic perfusion imaging of the myocardium was performed. Next, patients were given standard instructions for reporting chest pain symptoms during exercise. Patients then exercised using a Bruce or modified Bruce protocol, and were asked by the technician to report any symptoms at 2 minutes into each stage of exercise. At peak exercise, an additional 25.0 mCi of Tc-sestamibi was given, which was followed by tomographic perfusion imaging at least 30 minutes after exercise. Images for each patient were interpreted by a senior staff nuclear medicine physician who was unaware of patients’ symptom status during exercise as well as their health care utilization history. Images were classified as positive for ischemia if the image showed a reversible perfusion defect—a defect present after exercise but not before. The ECG taken during exercise was not used in the determining ischemia status because it has relatively low sensitivity and specificity. Symptoms during exercise that were considered to be angina- or cardiac-related were pain, tightness, pressure, or fullness in the chest; or discomfort radiating to the neck, jaw, arms, or upper back. Symptom reports that were limited to fatigue, dizziness, or shortness of
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
481
Table 1 Sociodemographics and health status variables of patients with silent or symptomatic myocardial ischemia
Gender, n (% male) Race, n (% white) Age, m (sd) Education, m (sd) Body mass index, m (sd) Current smoking, n (%) Hypertension, n (%) Hypercholesterolemia, n (%) Diabetes, n (%) Myocardial infarction, n (%)
Silent (n ⫽ 62)
Symptomatic (n ⫽ 33)
t or 2
p-Value
48 (77%) 45 (73%) 58.6 (10.6) 13.3 (2.7) 28.3 (4.8) 12 (19%) 33 (53%) 26 (42%) 11 (18%) 21 (34%)
25 (76%) 18 (55%) 56.4 (10.4) 12.5 (2.9) 30.5 (5.5) 4 (12%) 20 (61%) 15 (45%) 9 (27%) 15 (45%)
0.03 3.14 0.95 1.40 2.00 0.81 0.48 0.11 1.18 1.22
0.96 0.08 0.34 0.17 0.05 0.37 0.49 0.74 0.28 0.27
breath were not considered anginal or cardiac-related unless accompanied by at least one of the anginal symptoms. Health care utilization We obtained computerized health system out-patient utilization records for the 18 months prior to exercise testing, along with the department/setting visited and the primary diagnosis made at each visit. All visits were coded into one of four types of utilization based on the department/setting visited: emergency; primary care; specialty care; and cardiology. We tabulated the total number of visits in each setting for the 18-month period, after excluding any visit for which a cardiac- or chestpain-related diagnosis was given. Because all visits to cardiology had such a diagnosis, we did not eliminate cardiac- or chest-pain-related visits from this setting. Distributions of the number of visits to all settings were positively skewed. For example, for emergency care, about two thirds of the sample made no emergency visits, and over half of the rest made just one visit. One patient had five emergency contacts. This distribution is inappropriate for parametric analyses, so we recoded emergency care into a dichotomy—classifying each person has having either no emergency visits versus one or more visits. Visits to primary, specialty, and cardiology care had wider distributions, which allowed for parametric analyses, but these distributions also had some high outliers (i.e., patients with inordinately high utilization). We analyzed these variables in two ways. First, we treated these as continuous variables and analyzed them using parametric analyses, after recoding high outliers to 2 sd above the sample mean, in order to avoid biased analyses. We also dichotomized these variables—no visits versus one or more visits—in order to provide information about the absolute number of patients who sought each type of treatment. Results The final sample of 95 ischemic patients included 62 (65%) who had SMI, and 33 (35%) who reported angina
or other cardiac-related symptoms during exercise. Table 1 presents demographic and medical history data for both groups of patients. As can be seen, patients with SMI were nearly identical in gender distribution to patients with symptomatic ischemia. However, patients with SMI were significantly less obese and were somewhat more likely to be white than those with symptomatic ischemia. Although the two groups did not differ significantly with regard to age, education, or health history variables, the trends in this sample suggest that patients with SMI had fewer health problems than did symptomatic patients. Table 2 presents data on different types of health care visits for patients with SMI or symptomatic ischemia. First, the two groups were compared on utilization frequency via chi-square tests (for dichotomous data) or t-tests (for continuous data). As can be seen in Table 2, a significantly smaller percentage of the patients with SMI (21%) than with symptomatic ischemia (48.5%) sought emergency care during this period. In addition, patients with SMI were significantly less likely to seek primary care than were patients with symptomatic ischemia, and patients with SMI made fewer total primary care visits than did patients with symptomatic ischemia. The two groups did not differ in the likelihood or the frequency of visits to cardiology and other specialty settings. The observed differences in emergency and primary care utilization between the two groups could be accounted for by dissimilarities in demographic or health status variables. Thus, we again compared SMI patients with symptomatic patients on visits, this time controlling simultaneously for three demographic variables (age, education, race), and six health status variables (body mass index, smoking, hypertension, hypercholesterolemia, diabetes, history of myocardial infarction). These multivariate models (logistic regression or multiple regression), showed that patients with SMI remained less likely to use emergency care (Wald’s 2 ⫽ 3.97, p ⫽ 0.046), and remained less likely to use primary care (Wald’s 2 ⫽ 6.99, p ⫽ 0.008) and had a lower mean frequency of use of primary care [F(1,84) ⫽ 4.47, p ⫽
482
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
Table 2 Health care visits for noncardiac/chest pain-related diagnoses (except cardiology) over 18 months for patients with silent or symptomatic myocardial ischemia Type of visit Emergency care One or more visits [n (%)] Primary care One or more visits [n (%)] Frequency of visits [m (sd)] Specialty/other care One or more visits [n (%)] Frequency of visits [m (sd)] Cardiology One or more visits [n (%)] Frequency of visits [m (sd)]
Silent (n ⫽ 62)
Symptomatic (n ⫽ 33)
t or 2
p-Value
13 (21.0%)
16 (48.5%)
7.69
0.006
34 (54.8%) 2.09 (2.63)
27 (81.8%) 3.64 (2.95)
6.82 2.63
0.009 0.01
45 (72.6%) 3.70 (5.67)
25 (75.8%) 5.32 (6.90)
0.11 1.23
0.74 0.22
47 (75.6%) 1.76 (2.13)
21 (63.6%) 2.45 (3.05)
1.57 1.28
0.21 0.20
Data for “one or more visits” coded as 0 vs. 1⫹ visits. “Frequency of visits” had high outliers recoded to 2 sd above the sample mean.
0.037]; groups continued to show no difference in the presence/absence or mean frequency of visits to other specialty or cardiology settings. Finally, because some patients may have sought care outside of our health system of which we would be unaware, we examined utilization on the subset of patients (n ⫽ 74) insured by the managed care corporation affiliated with our system, for whom our utilization records were more complete. The 43 patients with SMI were less likely to have made an emergency visit (n ⫽ 9, 20.9%) than the 31 patients with symptomatic ischemia (n ⫽ 15, 48.4%; 2 ⫽ 6.20, p ⫽ 0.013). Likewise, patients with SMI made fewer primary care visits (m ⫽ 2.39, sd ⫽ 2.55) than the patients with symptomatic ischemia [m ⫽ 3.71, sd ⫽ 3.01; t(72) ⫽ 2.04, p ⫽ 0.045]. Finally, patients with SMI were marginally less likely to have made a primary care visit (n ⫽ 27, 62.8%) than patients with symptomatic ischemia (n ⫽ 25, 80.6%; 2 ⫽ 2.75, p ⫽ 0.097). Discussion Patients with silent myocardial ischemia, that is, who reported no angina or cardiac-related symptoms during exercise-induced myocardial ischemia, made fewer primary care visits and were less likely to seek emergency care during the prior 18 months than were patients who experienced angina in conjunction with ischemia. These differences in health care use were independent of various demographic and health status factors that might account for utilization differences, including the greater preponderance of whites and the lower obesity found among those with SMI. Importantly, these utilization differences were found for health care contacts associated with a range of somatic concerns unrelated to the heart or to chest pain. These findings are most consistent with the view that the failure to report angina during myocardial ischemia is not cardiac-specific, but rather is a generalized phe-
nomenon, applying to the whole person. The current findings add to a growing literature on the psychological correlates of SMI which support the view that the person rather than the event is silent [13–21]. We propose that the current behavioral findings, along with the many psychological constructs associated with SMI [13–21], reflect two basic personality dimensions. People with SMI tend to be low on the dimension of “openness”; that is, they focus their attention more on external than on internal events, including both psychological and somatic experiences. This is consistent with studies showing greater externally oriented, type A behavior, and possibly decreased pain sensitivity among SMI patients. Second, people with SMI tend to be low on the dimension of neuroticism or negative affectivity—they experience and report lower levels of aversive states such as depression, anxiety, and somatic symptoms, and they report higher levels of positive states such as marital satisfaction and an internal locus of control. Low levels of openness and negative affectivity may help account for reduced health care use among people with SMI. Studies have shown that health care utilization is precipitated less by the presence of disease than by elevated levels of somatic awareness, symptom concern, and emotional distress [32–35]. Because people with SMI appear to focus their attention externally and have relatively low levels of distress, they are less likely to seek medical care. The current study specifies this finding, however. In contrast to primary and emergency care, this study found no differences in specialty care utilization associated with SMI. We suspect that this is because primary and emergency care are typically initiated by the patient and, hence, influenced by patient characteristics such as openness and negative affectivity, whereas access to specialty care is often initiated and regulated by someone else, such as the primary care physician. Thus, patient characteristics may influence utilization more for some health care settings than others.
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
Although our view is that people with SMI are less likely to complain about their bodies and seek treatment, it should be recognized that there is another, logical interpretation. It is possible that patients with symptomatic ischemia have heightened sensitivity to bodily signals and, as a result, overly attend to their bodies and seek health care for various bodily problems. This proposal parallels observations of patients with panic attacks, who develop fear and somatic hypervigilance, which can lead to increased somatic complaints and treatment utilization, even for noncardiac or respiratory symptoms [36]. Unfortunately, studies that compare patients with silent and symptomatic myocardial ischemia cannot test whether silent ischemic patients are underreporters or symptomatic ischemic patients are overreporters. Resolving this dilemma will require either including a comparison group of patients without ischemia to determine “normative” utilization, or conducting a longitudinal study of patients at risk for ischemia to determine whether there are utilization differences before angina occurs, and whether utilization increases as a function of developing angina. There are several methodological limitations of our study. First, it should be noted that we chose to operationalize silent ischemia via reports during confirmed exercise-induced myocardial ischemia. This approach allowed greater standardization and control, yet reports of angina during treadmill testing often do not agree with in vivo (naturalistic) reports of angina; in particular, people lacking angina during laboratory exercise sometimes report angina in the natural environment. Although our finding of significant utilization differences associated with laboratory-based symptom reports indicates the validity of our approach, it is important to recognize that the patient groups we defined may be more similar in angina in the natural environment. Second, these two patient groups may differ in ways that we did not assess, especially aspects of their medical history such as the exact reasons that led them to be given exercise testing and their beliefs about their health. Although we have increasing knowledge about the general psychological characteristics of patients with SMI, the specific processes responsible for a lack of symptoms in SMI remain unclear. The “silence” of these patients may be due to a reduction in the generation of physiological sensations, a decreased perception of those sensations, a failure to interpret sensations as medical symptoms, and/or a reluctance to report symptoms to health professionals [22]. Further study will be needed to differentiate these processes and determine which contribute to SMI. Also, it remains possible that cardiacspecific factors, such as the size or location of cardiac ischemia, contribute to symptom silence in a subset of patients with SMI. The current data, however, provide support for an answer to the question posed by Barsky
483
et al. [22]. In general, it appears to be the person, rather than the event, that is silent. Acknowledgments This study was supported by a grant-in-aid from the American Heart Association of Michigan. We thank Debra Young, Jim Miller, and Darlene Calwell for their assistance in data collection. References [1] Deanfield JE, Maseri A, Selwyn AP, Chierchia S, Ribiero P, Krinkler S. Myocardial ischemia during daily life in patients with stable angina: its relation to symptoms and heart rate changes. Lancet 1983;ii:753. [2] Schang SJ, Pepine CJ. Transient asymptomatic ST segment depression during daily activity. Am J Cardiol 1977;39:396. [3] Hendler AL, Greyson ND, Robinson MG, Freeman MR. Patients with symptomatic ischemia have larger thallium perfusion abnormalities and more adverse prognosis than patients with silent ischemia. Can J Cardiol 1992;8:814–8. [4] Klein J, Chao SY, Berman DS, Rozanski A. Is “silent” myocardial ischaemia really as severe as symptomatic ischaemia? Circulation 1994;89:1958–66. [5] Nihoyannopoulos P, Marsonis A, Joshi J, Athanassopoulos G, Oakley CM. Magnitude of myocardial dysfunction is greater in painful than in painless myocardial ischemia: an exercise echocardiographic study. Am Coll Cardiol 1995;25:1507–12. [6] Droste C, Greenlee MW, Ruf G, Roskamm H. Localization of a coronary stenosis, left ventricular function, and pain perception during myocardial ischemia in patients with one-vessel disease. Cardiovasc Electrophysiol 1991;2(suppl.):S68–75. [7] Heller GV, Ahmed I, Tilkemeier PL, Barbour MM, Garber CE. Comparison of chest pain, electrocardiographic changes and Thallium-201 scintigraphy during varying exercise intensities in men with stable angina pectoris. Am J Cardiol 1991;68:569–74. [8] Reed DC, Stone PH, Shook TL, Shaw LA, Young P, Gibson RS. Discordance between radionuclide Doppler echo, angiographic and ambulatory ECG markers of ischemia. J Am Coll Cardiol 1989;13:2A. [9] Langer A, Freeman MR, Josse RG, Steiner G, Armstrong PW. Detection of silent myocardial ischemia in diabetes mellitus. Am J Cardiol 1991;67:1073–8. [10] Nesto RW, Phillips RT. Asymptomatic myocardial ischemia in diabetic patients. Am J Med 1986;80:40–7. [11] Rutherford JD, Braunwald E. Chronic ischemic heart disease. In: Braunwald E, editor. Heart Disease, 4th Ed. Philadelphia: Saunders, 1992, pp 1292–1364. [12] Deedwania PC, Jamner L, Carbajal E. Cigarette smoking increases risk of silent ischemia and alters cardiovascular reactivity during exercise. Circulation 1991;84:II-538. [13] Davies RF, Linden W, Habibi H, Klinke WP, Nadeau C, Phaneuf DC, Lepage S, Dessain P, Buttars JA. Relative importance of psychologic traits and severity of ischemia in causing angina during treadmill exercise. J Am Coll Cardiol 1993;21:331–6. [14] Droste C, Roskamm H. Experimental pain measurement in patients with asymptomatic myocardial ischemia. J Am Coll Cardiol 1983;1:940–5. [15] Freedland KE, Carney RM, Drone RJ, Smith LJ, Rich MW, Eisenkramer G, Fischer KC. Psychological factors in silent myocardial ischemia. Psychosom Med 1991;53:13–24. [16] Glazier JJ, Chierchia S, Brown MJ, Maseri A. Importance of generalized defective perception of painful stimuli as a cause of
484
[17]
[18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484 silent myocardial ischemia in chronic stable angina pectoris. Am J Cardiol 1986;58:667–72. Janne P, Reynaert C, Cassiers L, Huber W, deCoster P, Marchandise B, Kremer R. Psychological determinants of silent myocardial ischaemia. Eur Heart J 1987;8(suppl. G):125–9. Keefe FJ, Castell PJ, Blumenthal JA. Angina pectoris in type A and type B cardiac patients. Pain 1986;27:211–8. Sheps DS, Light KC, Bragdon EE. Relationship between chest pain, exercise endorphin response, and depression. Circulation 1989;80(suppl. 2):591. Siegel W, Mark D, Hlatky M, Pryor DB, Barefoot J, Williams R. Exercise-induced silent ischemia: effect of type A behavior on frequency and prognosis. Am J Cardiol 1989;64:1280–3. Torosian T, Lumley MA, Pickard SD, Ketterer MW. Silent versus symptomatic myocardial ischemia: the role of psychological and medical factors. Health Psychol 1997;16:123–30. Barsky AJ, Hochstrasser B, Coles A, Zisfein J, O’Donnell C, Eagle K. Silent myocardial ischemia: is the person or the event silent? JAMA 1990;264:1132–5. Carboni GP, Lahiri A, Cashman PMM, Raftery EB. Ambulatory heart rate and ST-segment depression during painful and silent myocardial ischemia in chronic stable angina pectoris. Am J Cardiol 1987;59:1029–34. Gasparetti CM, Burwell LR, Beller GA. Prevalence of and variables associated with silent myocardial ischemia on exercise thallium-201 stress testing. J Am Coll Cardiol 1990;16:115–23. Hecht HS, Shaw RE, Bruce T, Myler RK. Silent ischemia: evaluation by exercise and redistribution tomographic thallium-201 myocardial imaging. J Am Coll Cardiol 1989;14:895–900. Watson D, Clark L. Negative affectivity: the disposition to experience aversive emotional states. Psychol Bull 1984;96:465–90.
[27] Watson D, Pennebaker JW. Health complaints, stress, and distress: exploring the central role of negative affectivity. Psychol Rev 1989;96:234–54. [28] Lumley MA, Torosian T, Rowland LL, Ketterer MW, Pickard SD. Correlates of unrecognized acute myocardial infarction detected via perfusion imaging. Am J Cardiol 1997;79:1170–3. [29] Lumley MA, Torosian T, Ketterer MW, Pickard SD. Psychosocial factors related to non-cardiac chest pain during treadmill exercise. Psychosomatics 1997;38:230–8. [30] Lumley MA, Tomakowsky J, Torosian T. The relationship of alexithymia to subjective and biomedical measures of disease. Psychosomatics 1997;38:497–502. [31] Folstein MF, Folstein S, McHugh PR. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J Psychiatry Res 1975;12:189–98. [32] Barsky A, Wyshak G, Klerman G. Medical and psychiatric determinants of outpatient medical utilization. Med Care 1986;24: 548–60. [33] Cameron L, Leventhal EA, Leventhal H. Seeking medical care in response to symptoms and life stress. Psychosom Med 1995; 57:37–47. [34] Fitzgibbon ML, Stolley MR, Kirschenbaum DS. Obese people who seek treatment have different characteristics than those who do not seek treatment. Health Psychol 1993;12:342–5. [35] Stepanski E, Korshorek G, Zorick F, Glinn M, Roehrs T, Roth T. Characteristics of individuals who do or do not seek treatment for chronic insomnia. Psychosomatics 1989;30:421–7. [36] Zaubler TS, Katon W. Panic disorder in the general medical setting. J Psychosom Res 1998;44:25–42.
Decreased health care use among patients with silent myocardial ischemia: support for a generalized rather than cardiac-specific silence Mark Lumleya,*, Laurenn Rowlanda, Tracey Torosianb, Adam Banka, Mark Kettererb, Sol Pickardc a Department of Psychology, Wayne State University, 71 West Warren Avenue, Detroit, MI 48202, USA Department of Psychiatry, Henry Ford Health System, 2799 West Grand Boulevard, Detroit, MI 48202, USA c Division of Cardiovascular Medicine, Henry Ford Health System, 2799 West Grand Boulevard, Detroit, MI 48202, USA Received 30 September 1998; accepted 1 December 1999 b
Abstract Objective: The absence of angina among patients with silent myocardial ischemia (SMI) may be a cardiac phenomenon or may reflect a generalized lack of bodily awareness and symptom reporting. We tested the hypothesis that the silence is generalized, and, therefore, that patients with SMI would make fewer health care visits for noncardiac/chest-pain problems than patients with symptomatic ischemia. Methods: We counted all out-patient visits to our medical system for the prior 18 months for 95 patients who demonstrated ischemia during treadmill exercise testing and subsequent nuclear scanning: 62 of the patients had SMI during exercise, and 33 of
the patients had symptomatic ischemia. Results: Patients with SMI made were significantly less likely to have sought emergency care or primary care and had significantly fewer primary care visits than patients with symptomatic ischemia. Group differences remained after controlling for demographics and health status variables. The two groups did not differ on utilization of specialty care. Conclusion: The reduced use of emergency and primary care among patients with SMI suggests that they have a generalized rather than cardiac-specific reduction in somatic awareness and/or symptom reporting. 2000 Elsevier Science Inc. All rights reserved.
Keywords: Silent myocardial ischemia; Asymptomatic; Health care utilization; Chest pain; Somatic awareness
Introduction Coronary artery disease (CAD) results in myocardial ischemia when the oxygen supply is insufficient to meet myocardial demands. Ischemia may be accompanied by angina pectoris or other symptoms of pain and discomfort, but the ischemia is often silent or asymptomatic. A paradigm sometimes employed to study the correlates of silent myocardial ischemia (SMI) is to identify patients who exhibit ischemia during treadmill exercise testing, and then to compare the subsample of patients who report no symptoms during exercise-induced ischemia (i.e., who have SMI) with the subsample who report angina during ischemia (i.e., symptomatic ischemic patients). Studies have identified both biological and psychosocial correlates of SMI. Among the biological correlates, SMI has been linked to a lower heart rate at the onset * Corresponding author. Tel.: 313-577-2838; fax: 313-577-7636. E-mail address: [email protected] (M. Lumley).
of ischemia [1,2] and with smaller areas of ischemic tissue [3–5]. SMI also has been associated with specific regions of ischemic myocardium [6–8], diabetic neuropathy [9,10], the use of antianginal medications [11], and cigarette smoking [12]. Other studies have identified psychological correlates of SMI. Compared to patients with symptomatic ischemia, patients with SMI have been found to report less depression, anxiety, somatic awareness, illness seriousness, and sensitivity to pain and discomfort. They also report more marital satisfaction, type A behavior, and internal locus of control [13–20]. We have found that patients with SMI report a greater tendency to control their anger, and they prefer an externally oriented rather than an introspective style of thinking [21]. It remains unclear, however, whether the silence associated with SMI is limited to cardiac phenomena, or whether it reflects a more generalized somatic unawareness and underreporting of bodily symptoms. In a key review, Barsky and colleagues [22] raised this issue by
0022-3999/00/$ – see front matter 2000 Elsevier Science Inc. All rights reserved. PII: S0022-3999(00)00101-X
480
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
asking, “Is the person or the event silent?” Some of the cardiac correlates of SMI just noted (i.e., size and location of ischemic region, heart rate at ischemia onset, and use of antianginal medications) suggest that the silence is cardiac-specific—a function of cardiac tissue or innervation. In contrast, the findings associated with personality, attitudes, emotional style, and pain reporting suggest that the lack of symptoms is global or generalized—a function of the whole person. The available literature does not clearly answer the question of whether SMI is due to the person or the event. The cardiac correlates of SMI have been inconsistent. Some studies did not find lower heart rates at the onset of silent ischemia [21,23], and other studies found no associations between the size of ischemic myocardium and angina [21,24,25]. The psychological findings also have limitations. They have relied almost exclusively on self-reports via questionnaires, which, although purporting to assess different constructs, probably reflect a much smaller number of personality or emotional dimensions. For example, reports on many questionnaires are strongly influenced by a response tendency termed “negative affectivity,” which refers to a person’s disposition to experience and report negative aspects of oneself, including both emotional and physical symptoms [26,27]. Thus, those who experience SMI may simply be those who deny problems or have low levels of negative affectivity. We believe that a more valid test of the etiology of SMI requires measures that are not self-reported— measures such as behavioral indices. In this study, we examine the behavior of health care utilization. We reasoned that, if SMI is a generalized rather than cardiacspecific phenomenon, then patients with SMI would experience and complain of fewer bodily symptoms. This would result in patients with SMI being less likely to seek medical care for various somatic problems, particularly health problems unrelated to cardiac concerns and chest pain. In contrast, if SMI is cardiac-specific, then patients with SMI would be expected not to differ from symptomatic patients in their use of medical care for noncardiac/non–chest-pain problems. Thus, we compared patients who experienced SMI during treadmill exercise with patients who experienced symptomatic ischemia on the frequency with which they utilized various types of health care, including emergency, primary, specialty, and cardiology types of care. For each health care type (except cardiology), we excluded health care contacts that were related to cardiac problems or angina, thus ruling out the possibility that the decreased health care utilization was due solely to cardiac events and symptoms. Methods Participants Consecutive adults (n ⫽ 317) who had treadmill exercise testing with myocardial perfusion imaging were
asked to participate as they presented for testing; 275 (86.8%) agreed and provided informed consent. Of these, 17 were excluded due to cognitive impairment (n ⫽ 15) or illiteracy (n ⫽ 2). Of the remaining 258 patients, 112 (43.4%) were positive for exercise-induced cardiac ischemia (defined in what follows) and, of these patients, 95 (84.8%) were included because our records indicated that they received other out-patient care within our medical system during the prior 18 months. All of the patients in the final sample had private or public medical insurance, and most (n ⫽ 74, 78%) were insured by a managed care corporation affiliated with our health care system and which typically limited services to our system (thus providing us with patients’ complete utilization records). Most patients were referred for treadmill testing with nuclear scanning because they had experienced symptoms considered to be potentially indicative of coronary artery disease (CAD). Also, most patients, including those without symptoms, had other indications for testing, including elevated cardiac risk factors, abnormal or inconclusive resting electrocardiograms (ECGs), or presurgical clearance. Elsewhere, we reported on other characteristics from this cohort of patients, including: psychological correlates of SMI [21]; unrecognized myocardial infarction [28]; noncardiac chest pain [29]; and alexithymia [30]. Procedures Patients’ cognitive status was tested [31], they were interviewed and completed questionnaires for other studies, and were tested for myocardial ischemia. Before exercise, patients received 8.0 mCi of technetium-99msestamibi (Tc-setamibi), and baseline resting tomographic perfusion imaging of the myocardium was performed. Next, patients were given standard instructions for reporting chest pain symptoms during exercise. Patients then exercised using a Bruce or modified Bruce protocol, and were asked by the technician to report any symptoms at 2 minutes into each stage of exercise. At peak exercise, an additional 25.0 mCi of Tc-sestamibi was given, which was followed by tomographic perfusion imaging at least 30 minutes after exercise. Images for each patient were interpreted by a senior staff nuclear medicine physician who was unaware of patients’ symptom status during exercise as well as their health care utilization history. Images were classified as positive for ischemia if the image showed a reversible perfusion defect—a defect present after exercise but not before. The ECG taken during exercise was not used in the determining ischemia status because it has relatively low sensitivity and specificity. Symptoms during exercise that were considered to be angina- or cardiac-related were pain, tightness, pressure, or fullness in the chest; or discomfort radiating to the neck, jaw, arms, or upper back. Symptom reports that were limited to fatigue, dizziness, or shortness of
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
481
Table 1 Sociodemographics and health status variables of patients with silent or symptomatic myocardial ischemia
Gender, n (% male) Race, n (% white) Age, m (sd) Education, m (sd) Body mass index, m (sd) Current smoking, n (%) Hypertension, n (%) Hypercholesterolemia, n (%) Diabetes, n (%) Myocardial infarction, n (%)
Silent (n ⫽ 62)
Symptomatic (n ⫽ 33)
t or 2
p-Value
48 (77%) 45 (73%) 58.6 (10.6) 13.3 (2.7) 28.3 (4.8) 12 (19%) 33 (53%) 26 (42%) 11 (18%) 21 (34%)
25 (76%) 18 (55%) 56.4 (10.4) 12.5 (2.9) 30.5 (5.5) 4 (12%) 20 (61%) 15 (45%) 9 (27%) 15 (45%)
0.03 3.14 0.95 1.40 2.00 0.81 0.48 0.11 1.18 1.22
0.96 0.08 0.34 0.17 0.05 0.37 0.49 0.74 0.28 0.27
breath were not considered anginal or cardiac-related unless accompanied by at least one of the anginal symptoms. Health care utilization We obtained computerized health system out-patient utilization records for the 18 months prior to exercise testing, along with the department/setting visited and the primary diagnosis made at each visit. All visits were coded into one of four types of utilization based on the department/setting visited: emergency; primary care; specialty care; and cardiology. We tabulated the total number of visits in each setting for the 18-month period, after excluding any visit for which a cardiac- or chestpain-related diagnosis was given. Because all visits to cardiology had such a diagnosis, we did not eliminate cardiac- or chest-pain-related visits from this setting. Distributions of the number of visits to all settings were positively skewed. For example, for emergency care, about two thirds of the sample made no emergency visits, and over half of the rest made just one visit. One patient had five emergency contacts. This distribution is inappropriate for parametric analyses, so we recoded emergency care into a dichotomy—classifying each person has having either no emergency visits versus one or more visits. Visits to primary, specialty, and cardiology care had wider distributions, which allowed for parametric analyses, but these distributions also had some high outliers (i.e., patients with inordinately high utilization). We analyzed these variables in two ways. First, we treated these as continuous variables and analyzed them using parametric analyses, after recoding high outliers to 2 sd above the sample mean, in order to avoid biased analyses. We also dichotomized these variables—no visits versus one or more visits—in order to provide information about the absolute number of patients who sought each type of treatment. Results The final sample of 95 ischemic patients included 62 (65%) who had SMI, and 33 (35%) who reported angina
or other cardiac-related symptoms during exercise. Table 1 presents demographic and medical history data for both groups of patients. As can be seen, patients with SMI were nearly identical in gender distribution to patients with symptomatic ischemia. However, patients with SMI were significantly less obese and were somewhat more likely to be white than those with symptomatic ischemia. Although the two groups did not differ significantly with regard to age, education, or health history variables, the trends in this sample suggest that patients with SMI had fewer health problems than did symptomatic patients. Table 2 presents data on different types of health care visits for patients with SMI or symptomatic ischemia. First, the two groups were compared on utilization frequency via chi-square tests (for dichotomous data) or t-tests (for continuous data). As can be seen in Table 2, a significantly smaller percentage of the patients with SMI (21%) than with symptomatic ischemia (48.5%) sought emergency care during this period. In addition, patients with SMI were significantly less likely to seek primary care than were patients with symptomatic ischemia, and patients with SMI made fewer total primary care visits than did patients with symptomatic ischemia. The two groups did not differ in the likelihood or the frequency of visits to cardiology and other specialty settings. The observed differences in emergency and primary care utilization between the two groups could be accounted for by dissimilarities in demographic or health status variables. Thus, we again compared SMI patients with symptomatic patients on visits, this time controlling simultaneously for three demographic variables (age, education, race), and six health status variables (body mass index, smoking, hypertension, hypercholesterolemia, diabetes, history of myocardial infarction). These multivariate models (logistic regression or multiple regression), showed that patients with SMI remained less likely to use emergency care (Wald’s 2 ⫽ 3.97, p ⫽ 0.046), and remained less likely to use primary care (Wald’s 2 ⫽ 6.99, p ⫽ 0.008) and had a lower mean frequency of use of primary care [F(1,84) ⫽ 4.47, p ⫽
482
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
Table 2 Health care visits for noncardiac/chest pain-related diagnoses (except cardiology) over 18 months for patients with silent or symptomatic myocardial ischemia Type of visit Emergency care One or more visits [n (%)] Primary care One or more visits [n (%)] Frequency of visits [m (sd)] Specialty/other care One or more visits [n (%)] Frequency of visits [m (sd)] Cardiology One or more visits [n (%)] Frequency of visits [m (sd)]
Silent (n ⫽ 62)
Symptomatic (n ⫽ 33)
t or 2
p-Value
13 (21.0%)
16 (48.5%)
7.69
0.006
34 (54.8%) 2.09 (2.63)
27 (81.8%) 3.64 (2.95)
6.82 2.63
0.009 0.01
45 (72.6%) 3.70 (5.67)
25 (75.8%) 5.32 (6.90)
0.11 1.23
0.74 0.22
47 (75.6%) 1.76 (2.13)
21 (63.6%) 2.45 (3.05)
1.57 1.28
0.21 0.20
Data for “one or more visits” coded as 0 vs. 1⫹ visits. “Frequency of visits” had high outliers recoded to 2 sd above the sample mean.
0.037]; groups continued to show no difference in the presence/absence or mean frequency of visits to other specialty or cardiology settings. Finally, because some patients may have sought care outside of our health system of which we would be unaware, we examined utilization on the subset of patients (n ⫽ 74) insured by the managed care corporation affiliated with our system, for whom our utilization records were more complete. The 43 patients with SMI were less likely to have made an emergency visit (n ⫽ 9, 20.9%) than the 31 patients with symptomatic ischemia (n ⫽ 15, 48.4%; 2 ⫽ 6.20, p ⫽ 0.013). Likewise, patients with SMI made fewer primary care visits (m ⫽ 2.39, sd ⫽ 2.55) than the patients with symptomatic ischemia [m ⫽ 3.71, sd ⫽ 3.01; t(72) ⫽ 2.04, p ⫽ 0.045]. Finally, patients with SMI were marginally less likely to have made a primary care visit (n ⫽ 27, 62.8%) than patients with symptomatic ischemia (n ⫽ 25, 80.6%; 2 ⫽ 2.75, p ⫽ 0.097). Discussion Patients with silent myocardial ischemia, that is, who reported no angina or cardiac-related symptoms during exercise-induced myocardial ischemia, made fewer primary care visits and were less likely to seek emergency care during the prior 18 months than were patients who experienced angina in conjunction with ischemia. These differences in health care use were independent of various demographic and health status factors that might account for utilization differences, including the greater preponderance of whites and the lower obesity found among those with SMI. Importantly, these utilization differences were found for health care contacts associated with a range of somatic concerns unrelated to the heart or to chest pain. These findings are most consistent with the view that the failure to report angina during myocardial ischemia is not cardiac-specific, but rather is a generalized phe-
nomenon, applying to the whole person. The current findings add to a growing literature on the psychological correlates of SMI which support the view that the person rather than the event is silent [13–21]. We propose that the current behavioral findings, along with the many psychological constructs associated with SMI [13–21], reflect two basic personality dimensions. People with SMI tend to be low on the dimension of “openness”; that is, they focus their attention more on external than on internal events, including both psychological and somatic experiences. This is consistent with studies showing greater externally oriented, type A behavior, and possibly decreased pain sensitivity among SMI patients. Second, people with SMI tend to be low on the dimension of neuroticism or negative affectivity—they experience and report lower levels of aversive states such as depression, anxiety, and somatic symptoms, and they report higher levels of positive states such as marital satisfaction and an internal locus of control. Low levels of openness and negative affectivity may help account for reduced health care use among people with SMI. Studies have shown that health care utilization is precipitated less by the presence of disease than by elevated levels of somatic awareness, symptom concern, and emotional distress [32–35]. Because people with SMI appear to focus their attention externally and have relatively low levels of distress, they are less likely to seek medical care. The current study specifies this finding, however. In contrast to primary and emergency care, this study found no differences in specialty care utilization associated with SMI. We suspect that this is because primary and emergency care are typically initiated by the patient and, hence, influenced by patient characteristics such as openness and negative affectivity, whereas access to specialty care is often initiated and regulated by someone else, such as the primary care physician. Thus, patient characteristics may influence utilization more for some health care settings than others.
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484
Although our view is that people with SMI are less likely to complain about their bodies and seek treatment, it should be recognized that there is another, logical interpretation. It is possible that patients with symptomatic ischemia have heightened sensitivity to bodily signals and, as a result, overly attend to their bodies and seek health care for various bodily problems. This proposal parallels observations of patients with panic attacks, who develop fear and somatic hypervigilance, which can lead to increased somatic complaints and treatment utilization, even for noncardiac or respiratory symptoms [36]. Unfortunately, studies that compare patients with silent and symptomatic myocardial ischemia cannot test whether silent ischemic patients are underreporters or symptomatic ischemic patients are overreporters. Resolving this dilemma will require either including a comparison group of patients without ischemia to determine “normative” utilization, or conducting a longitudinal study of patients at risk for ischemia to determine whether there are utilization differences before angina occurs, and whether utilization increases as a function of developing angina. There are several methodological limitations of our study. First, it should be noted that we chose to operationalize silent ischemia via reports during confirmed exercise-induced myocardial ischemia. This approach allowed greater standardization and control, yet reports of angina during treadmill testing often do not agree with in vivo (naturalistic) reports of angina; in particular, people lacking angina during laboratory exercise sometimes report angina in the natural environment. Although our finding of significant utilization differences associated with laboratory-based symptom reports indicates the validity of our approach, it is important to recognize that the patient groups we defined may be more similar in angina in the natural environment. Second, these two patient groups may differ in ways that we did not assess, especially aspects of their medical history such as the exact reasons that led them to be given exercise testing and their beliefs about their health. Although we have increasing knowledge about the general psychological characteristics of patients with SMI, the specific processes responsible for a lack of symptoms in SMI remain unclear. The “silence” of these patients may be due to a reduction in the generation of physiological sensations, a decreased perception of those sensations, a failure to interpret sensations as medical symptoms, and/or a reluctance to report symptoms to health professionals [22]. Further study will be needed to differentiate these processes and determine which contribute to SMI. Also, it remains possible that cardiacspecific factors, such as the size or location of cardiac ischemia, contribute to symptom silence in a subset of patients with SMI. The current data, however, provide support for an answer to the question posed by Barsky
483
et al. [22]. In general, it appears to be the person, rather than the event, that is silent. Acknowledgments This study was supported by a grant-in-aid from the American Heart Association of Michigan. We thank Debra Young, Jim Miller, and Darlene Calwell for their assistance in data collection. References [1] Deanfield JE, Maseri A, Selwyn AP, Chierchia S, Ribiero P, Krinkler S. Myocardial ischemia during daily life in patients with stable angina: its relation to symptoms and heart rate changes. Lancet 1983;ii:753. [2] Schang SJ, Pepine CJ. Transient asymptomatic ST segment depression during daily activity. Am J Cardiol 1977;39:396. [3] Hendler AL, Greyson ND, Robinson MG, Freeman MR. Patients with symptomatic ischemia have larger thallium perfusion abnormalities and more adverse prognosis than patients with silent ischemia. Can J Cardiol 1992;8:814–8. [4] Klein J, Chao SY, Berman DS, Rozanski A. Is “silent” myocardial ischaemia really as severe as symptomatic ischaemia? Circulation 1994;89:1958–66. [5] Nihoyannopoulos P, Marsonis A, Joshi J, Athanassopoulos G, Oakley CM. Magnitude of myocardial dysfunction is greater in painful than in painless myocardial ischemia: an exercise echocardiographic study. Am Coll Cardiol 1995;25:1507–12. [6] Droste C, Greenlee MW, Ruf G, Roskamm H. Localization of a coronary stenosis, left ventricular function, and pain perception during myocardial ischemia in patients with one-vessel disease. Cardiovasc Electrophysiol 1991;2(suppl.):S68–75. [7] Heller GV, Ahmed I, Tilkemeier PL, Barbour MM, Garber CE. Comparison of chest pain, electrocardiographic changes and Thallium-201 scintigraphy during varying exercise intensities in men with stable angina pectoris. Am J Cardiol 1991;68:569–74. [8] Reed DC, Stone PH, Shook TL, Shaw LA, Young P, Gibson RS. Discordance between radionuclide Doppler echo, angiographic and ambulatory ECG markers of ischemia. J Am Coll Cardiol 1989;13:2A. [9] Langer A, Freeman MR, Josse RG, Steiner G, Armstrong PW. Detection of silent myocardial ischemia in diabetes mellitus. Am J Cardiol 1991;67:1073–8. [10] Nesto RW, Phillips RT. Asymptomatic myocardial ischemia in diabetic patients. Am J Med 1986;80:40–7. [11] Rutherford JD, Braunwald E. Chronic ischemic heart disease. In: Braunwald E, editor. Heart Disease, 4th Ed. Philadelphia: Saunders, 1992, pp 1292–1364. [12] Deedwania PC, Jamner L, Carbajal E. Cigarette smoking increases risk of silent ischemia and alters cardiovascular reactivity during exercise. Circulation 1991;84:II-538. [13] Davies RF, Linden W, Habibi H, Klinke WP, Nadeau C, Phaneuf DC, Lepage S, Dessain P, Buttars JA. Relative importance of psychologic traits and severity of ischemia in causing angina during treadmill exercise. J Am Coll Cardiol 1993;21:331–6. [14] Droste C, Roskamm H. Experimental pain measurement in patients with asymptomatic myocardial ischemia. J Am Coll Cardiol 1983;1:940–5. [15] Freedland KE, Carney RM, Drone RJ, Smith LJ, Rich MW, Eisenkramer G, Fischer KC. Psychological factors in silent myocardial ischemia. Psychosom Med 1991;53:13–24. [16] Glazier JJ, Chierchia S, Brown MJ, Maseri A. Importance of generalized defective perception of painful stimuli as a cause of
484
[17]
[18] [19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
M. Lumley et al. / Journal of Psychosomatic Research 48 (2000) 479–484 silent myocardial ischemia in chronic stable angina pectoris. Am J Cardiol 1986;58:667–72. Janne P, Reynaert C, Cassiers L, Huber W, deCoster P, Marchandise B, Kremer R. Psychological determinants of silent myocardial ischaemia. Eur Heart J 1987;8(suppl. G):125–9. Keefe FJ, Castell PJ, Blumenthal JA. Angina pectoris in type A and type B cardiac patients. Pain 1986;27:211–8. Sheps DS, Light KC, Bragdon EE. Relationship between chest pain, exercise endorphin response, and depression. Circulation 1989;80(suppl. 2):591. Siegel W, Mark D, Hlatky M, Pryor DB, Barefoot J, Williams R. Exercise-induced silent ischemia: effect of type A behavior on frequency and prognosis. Am J Cardiol 1989;64:1280–3. Torosian T, Lumley MA, Pickard SD, Ketterer MW. Silent versus symptomatic myocardial ischemia: the role of psychological and medical factors. Health Psychol 1997;16:123–30. Barsky AJ, Hochstrasser B, Coles A, Zisfein J, O’Donnell C, Eagle K. Silent myocardial ischemia: is the person or the event silent? JAMA 1990;264:1132–5. Carboni GP, Lahiri A, Cashman PMM, Raftery EB. Ambulatory heart rate and ST-segment depression during painful and silent myocardial ischemia in chronic stable angina pectoris. Am J Cardiol 1987;59:1029–34. Gasparetti CM, Burwell LR, Beller GA. Prevalence of and variables associated with silent myocardial ischemia on exercise thallium-201 stress testing. J Am Coll Cardiol 1990;16:115–23. Hecht HS, Shaw RE, Bruce T, Myler RK. Silent ischemia: evaluation by exercise and redistribution tomographic thallium-201 myocardial imaging. J Am Coll Cardiol 1989;14:895–900. Watson D, Clark L. Negative affectivity: the disposition to experience aversive emotional states. Psychol Bull 1984;96:465–90.
[27] Watson D, Pennebaker JW. Health complaints, stress, and distress: exploring the central role of negative affectivity. Psychol Rev 1989;96:234–54. [28] Lumley MA, Torosian T, Rowland LL, Ketterer MW, Pickard SD. Correlates of unrecognized acute myocardial infarction detected via perfusion imaging. Am J Cardiol 1997;79:1170–3. [29] Lumley MA, Torosian T, Ketterer MW, Pickard SD. Psychosocial factors related to non-cardiac chest pain during treadmill exercise. Psychosomatics 1997;38:230–8. [30] Lumley MA, Tomakowsky J, Torosian T. The relationship of alexithymia to subjective and biomedical measures of disease. Psychosomatics 1997;38:497–502. [31] Folstein MF, Folstein S, McHugh PR. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J Psychiatry Res 1975;12:189–98. [32] Barsky A, Wyshak G, Klerman G. Medical and psychiatric determinants of outpatient medical utilization. Med Care 1986;24: 548–60. [33] Cameron L, Leventhal EA, Leventhal H. Seeking medical care in response to symptoms and life stress. Psychosom Med 1995; 57:37–47. [34] Fitzgibbon ML, Stolley MR, Kirschenbaum DS. Obese people who seek treatment have different characteristics than those who do not seek treatment. Health Psychol 1993;12:342–5. [35] Stepanski E, Korshorek G, Zorick F, Glinn M, Roehrs T, Roth T. Characteristics of individuals who do or do not seek treatment for chronic insomnia. Psychosomatics 1989;30:421–7. [36] Zaubler TS, Katon W. Panic disorder in the general medical setting. J Psychosom Res 1998;44:25–42.